← Back to the Hub
Preset
kimi-k3-coding
kimi-k3-coding
Kimi K3 long-horizon coding harness with 1M context, max reasoning, native tools, and cache-friendly extended history.
builtinOpen sourceReadyContext replay sessions
Specification
| Category | Model and task presets |
|---|---|
| Runtime | builtin |
| Interface | TUI and CLI |
| Source | built-in-template |
| Licence | Apache-2.0 |
| Provider | moonshot |
| Model | kimi-k3 |
| Continuity | Context replay |
| Support | Ready |
| Based on | kimi-k3-coding |
Runtime behaviour
Capabilities
single workflowParallel tool callsStructured run events
Tools
read_filewrite_filelist_directoryedit_fileinsert_textpatchmulti_editgrepgloblocal_code_searchrepo_searchcode_searchbashtodo_writetodo_readPolicies and ownership
- +Sandbox: local
- +Approvals: balanced
- +Shell: allowed
- +Network: blocked
- +Checks: enabled
Get started
Inside SuperQode
$ :harness switch kimi-k3-coding$ :harness show kimi-k3-coding$ :harness statusFrom your shell
$ superqode harness run kimi-k3-coding "Describe your task"$ superqode harness show kimi-k3-codingEvaluate and optimize
Measure
$ superqode harness test --spec harness.yaml$ superqode harness eval --spec harness.yaml --tasks eval-tasks.yaml$ superqode harness bench --manifest harnessbench.yamlImprove
$ superqode harness optimize-omni --spec harness.yaml --tasks eval-tasks.yaml$ superqode harness optimize --spec harness.yaml --tasks eval-tasks.yaml$ superqode harness promote stage